Income Prediction using Adult Census Data – Python, Scikit-learn, Pandas

Income prediction is one the biggest challenge to foresee the economy. I collected the Adult census data and researched the useful parameters that helps in the income prediction.

The data pre-processing phase is the toughest part since revolves around data gathering and cleaning, performing exploratory data analysis.

I performed classification task to predict whether the income exceeds 50k USD per year using different classification models(Machine learning Algorithms) including cost sensitive and data sampling algorithms on 2 million records of data. I used scikit-learn and pandas to deal with data.

Achieved 95% test data accuracy through XG Boost model and by using ROC-AUC evaluation metric

You can find the python code and Sample data here: